Graphical Models

Adams, Roy, Nazir Saleheen, Edison Thomaz, Abhinav Parate, Santosh Kumar, and Benjamin Marlin. "Hierarchical Span-Based Conditional Random Fields for Labeling and Segmenting Events in Wearable Sensor Data Streams." International Conference on Machine Learning. 2016. Abstracticml2016_hns.pdf

The field of mobile health (mHealth) has the potential to yield new insights into health and behavior through the analysis of continuously recorded data from wearable health and activity sensors. In this paper, we present a hierarchical span-based conditional random field model for the key problem of jointly detecting discrete events in such sensor data streams and segmenting these events into high-level activity sessions. Our model includes higher-order cardinality factors and inter-event duration factors to capture domain-specific structure in the label space. We show that our model supports exact MAP inference in quadratic time via dynamic programming, which we leverage to perform learning in the structured support vector machine framework. We apply the model to the problems of smoking and eating detection using four real data sets. Our results show statistically significant improvements in segmentation performance at the p=0.005 level relative to a hierarchical pairwise CRF.

Nguyen, Thai, Roy J. Adams, Annamalai Natarajan, and Benjamin M. Marlin. "Parsing Wireless Electrocardiogram Signals with Context Free Grammar Conditional Random Fields." IEEE Wireless Health. 2016. Abstractnguyen-wh2016.pdf

Recent advances in wearable sensor technology have made it possible to simultaneously collect multiple streams of physiological and context data from individuals as they go about their daily activities in natural environments. However, extracting reliable higher-level inferences from these raw data streams remains a key data analysis challenge. In this paper, we focus on the specific case of the analysis of data from wireless electrocardiogram (ECG) sensors. We present a new robust probabilistic approach to ECG morphology extraction using conditional random field context free grammar models, which have traditionally been applied to parsing problems in natural language processing. We introduce a robust context free grammar for parsing noisy ECG data, and show significantly improved performance on the ECG morphological labeling task.

Adams, Roy J., Edison Thomaz, and Benjamin M. Marlin. "Hierarchical Nested CRFs for Segmentation and Labeling of Physiological Time Series." NIPS Workshop on Machine Learning in Healthcare. 2015. Abstractadams-nips-heath2015.pdf

In this paper, we address the problem of nested hierarchical segmentation
and labeling of time series data. We present a hierarchical
span-based conditional random field framework for this problem that
leverages higher-order factors to enforce the nesting constraints. The framework can
incorporate a variety of additional factors including higher order cardinality
factors. This research is motivated by hierarchical activity recognition problems
in the field of mobile Health (mHealth). We show that the specific model of interest in the mHealth setting supports exact MAP inference in quadratic time. Learning is accomplished in the structured support vector machine framework. We show positive results on real and synthetic data sets.

Natarajan, Annamalai, Edward Gaiser, Gustavo Angarita, Robert Malison, Deepak Ganesan, and Benjamin Marlin. "Conditional Random Fields for Morphological Analysis of Wireless ECG Signals." Proceedings of the 5th Annual conference on Bioinformatics, Computational Biology and Health Informatics. Newport Beach, CA: ACM, 2014. Abstractcrf_bcb2014.pdf

Thanks to advances in mobile sensing technologies, it has recently become practical to deploy wireless electrocardiograph sensors for continuous recording of ECG signals. This capability has diverse applications in the study of human health and behavior, but to realize its full potential, new computational tools are required to effectively deal with the uncertainty that results from the noisy and highly non-stationary signals collected using these devices. In this work, we present a novel approach to the problem of extracting the morphological structure of ECG signals based on the use of dynamically structured conditional random field (CRF) models. We apply this framework to the problem of extracting morphological structure from wireless ECG sensor data collected in a lab-based study of habituated cocaine users. Our results show that the proposed CRF-based approach significantly out-performs independent prediction models using the same features, as well as a widely cited open source toolkit.

Parate, Abhinav, Meng-Chieh Chiu, Deepak Ganesan, and Benjamin M. Marlin. "Leveraging graphical models to improve accuracy and reduce privacy risks of mobile sensing." MobiSys. 2013. 83-96. Abstractsensing_mobisys13_paper.pdf

The proliferation of sensors on mobile phones and wearables has led to a plethora of context classifiers designed to sense the individual's context. We argue that a key missing piece in mobile inference is a layer that fuses the outputs of several classifiers to learn deeper insights into an individual's habitual patterns and associated correlations between contexts, thereby enabling new systems optimizations and opportunities. In this paper, we design CQue, a dynamic bayesian network that operates over classifiers for individual contexts, observes relations across these outputs across time, and identifies opportunities for improving energy-efficiency and accuracy by taking advantage of relations. In addition, such a layer provides insights into privacy leakage that might occur when seemingly innocuous user context revealed to different applications on a phone may be combined to reveal more information than originally intended. In terms of system architecture, our key contribution is a clean separation between the detection layer and the fusion layer, enabling classifiers to solely focus on detecting the context, and leverage temporal smoothing and fusion mechanisms to further boost performance by just connecting to our higher-level inference engine. To applications and users, CQue provides a query interface, allowing a) applications to obtain more accurate context results while remaining agnostic of what classifiers/sensors are used and when, and b) users to specify what contexts they wish to keep private, and only allow information that has low leakage with the private context to be revealed. We implemented CQue in Android, and our results show that CQue can i) improve activity classification accuracy up to 42%, ii) reduce energy consumption in classifying social, location and activity contexts with high accuracy(>90%) by reducing the number of required classifiers by at least 33%, and iii) effectively detect and suppress contexts that reveal private information.

Marlin, Benjamin M., and Nando de Freitas. "Asymptotic Efficiency of Deterministic Estimators for Discrete Energy-Based Models: Ratio Matching and Pseudolikelihood." UAI. 2011. 497-505. Abstractestimators_uai11_paper.pdf

Standard maximum likelihood estimation cannot be applied to discrete energy-based models in the general case because the computation of exact model probabilities is intractable. Recent research has seen the proposal of several new estimators designed specifically to overcome this intractability, but virtually nothing is known about their theoretical properties. In this paper, we present a generalized estimator that unifies many of the classical and recently proposed estimators. We use results from the standard asymptotic theory for M-estimators to derive a generic expression for the asymptotic covariance matrix of our generalized estimator. We apply these results to study the relative statistical efficiency of classical pseudolikelihood and the recently-proposed ratio matching estimator.

Duvenaud, David K., Benjamin M. Marlin, and Kevin P. Murphy. "Multiscale Conditional Random Fields for Semi-supervised Labeling and Classification." CRV. 2011. 371-378. Abstractmultiscale_crv11_paper.pdf

Motivated by the abundance of images labeled only by their captions, we construct tree-structured multiscale conditional random fields capable of performing semi supervised learning. We show that such caption-only data can in fact increase pixel-level accuracy at test time. In addition, we compare two kinds of tree: the standard one with pair wise potentials, and one based on noisy-or potentials, which better matches the semantics of the recursive partitioning used to create the tree.

Swersky, Kevin, Marc'Aurelio Ranzato, David Buchman, Benjamin M. Marlin, and Nando de Freitas. "On Autoencoders and Score Matching for Energy Based Models." ICML. 2011. 1201-1208. Abstractautoencoders_icml11_paper.pdf

We consider estimation methods for the class of continuous-data energy based models (EBMs). Our main result shows that estimating the parameters of an EBM using score matching when the conditional distribution over the visible units is Gaussian corresponds to training a particular form of regularized autoencoder. We show how different Gaussian EBMs lead to different autoencoder architectures, providing deep links between these two families of models. We compare the score matching estimator for the mPoT model, a particular Gaussian EBM, to several other training methods on a variety of tasks including image denoising and unsupervised feature extraction. We show that the regularization function induced by score matching leads to superior classification performance relative to a standard autoencoder. We also show that score matching yields classification results that are indistinguishable from better-known stochastic approximation maximum likelihood estimators.